a comparative study of croatian and hungarian 8th graders ... · croatian is a slavic language,...

1

Jelena Mihaljević Djigunović

English Department, University of Zagreb,

Ivana Lucica 3, Zagreb, 10 000, Croatia

[email protected]

Marianne Nikolov

(corresponding author)

Department of English Applied Linguistics, University of Pécs,

Pécs, 7624, Ifjúság u. 6. Hungary

[email protected]

István Ottó

Mottologic,

Kaposvár, 7400, Petőfi u. 5

[email protected]

A comparative study of Croatian and Hungarian 8th graders'

performances in English

*The research project was supported by a grant from the CERGE-EI Foundation under a

program of the Global Development Network. Additional funds for grantees in the

Balkan countries have been provided by the Austrian Government through WIIW,

Vienna. All opinions expressed are those of the authors and have not been endorsed by

CERGE-EI, WIIW, or the GDN.

2

A comparative study of Croatian and Hungarian 8th graders'

performances in English

Abstract

This comparative research aims to provide insights into Croatian and Hungarian 8th

graders’ performances in English as a foreign language (EFL). A total of 717 14-year-old

students participated in the study in two neighboring regions and towns of Croatia and

Hungary. The study examines how Croatian and Hungarian 8th graders’ performances

compare on EFL tests; how they compare by groups, within groups, by length of

instruction, weekly classes, and size of group; and the relationship between Croatian and

Hungarian students’ achievements on tests in their mother tongue (L1) and EFL. The

findings show that Croatian students performed significantly better on the EFL

proficiency tests than their Hungarian counterparts; larger differences have been found

between groups’ performances in Hungary than in Croatia, whereas no significant

differences characterize within-group variations. Learners who started EFL earlier tend to

achieve higher scores than later starters, whereas the findings on group size and weekly

classes are more controversial.

3

Introduction

The present study is part of a larger cross-country project carried out in Hungary

and Croatia with the same constructs, measuring instruments, and procedures. A total of

717 14-year-old students participated in the study in two neighboring regions and towns

of Croatia and Hungary. The project aims (1) to find out what Croatian and Hungarian

learners of English as a foreign language (EFL) can do at the end of year 8; (2) to

compare learners’ performances on EFL proficiency tests along some educational

variables. The present study examines how Croatian and Hungarian 8th graders’

performances compare on EFL tests; how they compare by groups, within groups, by

length of instruction, weekly classes, and size of group; and the relationship between

Croatian and Hungarian students’ achievements on tests in their mother tongue (L1) and

EFL.

Background to study

Theoretical background

The most relevant theoretical underpinnings of the study relate to several second

language acquisition (SLA) research areas: (1) how young learners develop (Cameron,

2001; Curtain & Pesola, 1994; Moon, 2000; Nikolov, 2002a) in foreign language

classrooms, where they have limited access to the target language and it is one of the

school subjects. As Johnstone (2000, p. 128) pointed out, there has been a lack of studies

inquiring into how primary-school learners’ achievements and other variables interact in

foreign language contexts. (2) the assessment of communicative language ability

(Bachman & Palmer, 1996) and communicative language competence (Common

European framework of reference for languages: learning, teaching, assessment, 2001)

(CEFR); and (3) the interdependence of L1 and L2 in emerging bilingualism (Baker,

1996; Cummins, 1991, 2000). The term “common underlying proficiency” (Cummins,

2000, p. 38) refers to cognitive academic proficiency underlying performance in both

languages and the interdependence hypothesis states that proficiency transfer from one

language to the other will occur provided there is adequate exposure to L2 in school as

well as in the environment and adequate motivation to learn it. This means that the

linguistic and literacy knowledge and skills students have learned in their L1 will bear on

4

the learning of academic knowledge and skills in L2 involving, for example, knowledge

of how to approach a text and background knowledge of the world (Cummins, 2000, p.

190). In bilingual educational contexts research has shown that cross-lingual influence

can operate in both directions (e.g., Verhoeven, 1991). A number of studies have

supported the interdependence hypothesis across a variety of languages, though lower

correlations have been found between linguistically distant languages (e. g., Chinese–

English, Turkish–Dutch) than for languages that are relatively close to each other (e. g.,

English–French) (Baker, 1996; Cummins, 2000; Genesee, 1979).

A related but different hypothesis has been put forth in reading research

conceptualizing the complex relationship between L1 and L2, and within L2, whether

reading in L2 is a language problem or a reading problem (Alderson, 1984). Three types

of data would be necessary to test this threshold hypothesis: reading ability in L1 and L2

and proficiency in L2 of the same individuals. Research in this area has resulted in

contradictory findings (e. g. Lee & Schallert, 1997), as the separation of L2 proficiency

and L2 reading skills is problematic. It is clear from the literature that L2 knowledge and

reading knowledge interact: “the evidence is that, in second-language reading, knowledge

of the second language is a more important factor than the first-language reading

abilities” (Alderson, 2000, p. 23). However, testing the threshold hypothesis has so far

proved impossible. As to the extent to which these factors interact, Bernhardt (2000;

2003 cited in Brantmeier, 2004, p. 53) found that 50% of L2 reading was accounted for

by L1 literacy (20%) and L2 knowledge (30%), but the remaining 50% of variance was

unexplained.

The two educational contexts

Although Croatia and Hungary are neighbors and share many educational aspects

and problems, important characteristics are different in their language education. In our

view, the most significant areas relate to the following: (1) language levels and language

learning traditions, (2) language teacher education, (3) teacher supply in schools, (4)

quality of teaching, (5) the amount and quality of exposure to English available to

students and teachers outside the classroom, and (6) the relationship between these

factors and learners’ socioeconomic status. Also, (7) linguistically, the official languages

5

spoken by the vast majority in the two countries belong to different language families:

Croatian is a Slavic language, whereas Hungarian, a Finno-Ugric language, does not

belong to the Indo-European language family.

Croatia has gone through major changes in FL education policies in recent

history. For political reasons, after 1948 Russian stopped being the compulsory FL in

schools in Croatia (as part of former Yugoslavia); the repertoire of the first FL was

enlarged to include also French, German and English. In the 1960s German became the

preferred FL due to economic reasons: large numbers of Croats went to live in Germany

and Austria as guest workers. In the ‘70s the popularity of German was reinforced as

Croatia came to be recognized as an attractive tourist country and the majority of tourists

came from Germany and Austria. English was making its way slowly during these

decades and soon became a popular FL. As Croatia was more open to foreign influences,

cultures and products than other countries from behind the Iron Curtain, it was easy to

keep in touch with the trends in the world in all areas of life. American and British films,

music, and magazines were available. With English becoming an international language,

it turned into the most popular FL in the country. For a time, Croatian officials struggled

to keep different foreign languages on offer in schools, sometimes insisting that the four

languages (English, French, German and Russian, or Italian in regions bordering on Italy)

be secured within a set of local schools. This idea was, however, soon abandoned due to

the pressure of parents and pupils who insisted on English. Those learners who could not

get English started with another FL but were offered English either later or as an optional

school subject.

The majority of Croatians can use foreign languages, most importantly English, to

communicate with foreigners for personal or professional purposes. Unfortunately, this

statement is based on anecdotal evidence, as no studies have been conducted on Croats’

command of FLs and census data are not available either. It is only in 2003 that

systematic data collection on Croats’ FL competences was undertaken but results have

not been published.

In Hungary German used to be perceived as the most useful foreign language for

historical and economic reasons and access to it was the privilege of the exceptional few.

For a long time, however, tourism and business relations were on a significantly lower

6

level than in Croatia. Thus, the Hungarian population’s levels of proficiency in modern

languages have been extremely low: the ratio of citizens claiming to know at least one

foreign language, according to data in the most recent National Census (2002), is 19

percent, lagging behind all nations in the European Union (Europeans and languages: A

Eurobarometer special survey, 2001). These data are based on self-reports; actual levels

may be even lower. Recently, English as a world language has become the most

prestigious and useful foreign language, whereas German has been losing some of its

appeal in schools and in society (Csapó, 2001; Halász & Lannert, 2003; Nikolov, 1999a).

One of the reasons why language levels are low is that in Hungary Russian was

the mandatory foreign language for over four decades and although all students studied it

from age ten, few achieved a functional level as Russian was not taught for

communicative purposes. In addition to Russian, students in grammar schools were

taught another modern foreign language. The situation changed overnight in 1989: since

then, students may choose a foreign language they wish to study.

As for foreign language teacher education traditions, the two countries contrast in

important ways. Croatia, generally speaking, has had qualified language teachers for a

long time. It was only after English started to gain in popularity around the 1950s when

teachers of German and, especially, French had to be retrained in fast-track programs to

meet the growing demand due to increasing numbers of students enrolled in English

programs.

In Hungary, the 1990s witnessed major innovations in language teacher

education. Because of the serious shortage of modern language teachers, Russian teachers

and other majors were retrained on the job as teachers of English and German to meet

demands. The majority of retrained teachers kept their tenured posts in state schools; as a

result, 65 percent of teachers of modern languages in primary schools are retrained

graduates (Halász & Lannert, 1998, p. 273). During the first half of the 1990s

international organizations contributed to the teaching of foreign languages and training

teachers. However, despite the fact that the number of English and German majors in

both fast-track and traditional programs increased significantly, even today ten percent of

English and German teachers are unqualified (Halász & Lannert, 2003), as young

language major graduates find better paid jobs on the market.

7

In Croatia university-level teaching degrees have been required for decades.

Teachers’ overall proficiency is high because of a longer tradition in language teacher

education, more exposure to authentic language in the media and in everyday life, and

stronger instrumental motivation to know languages because of tourism and business

opportunities. In Croatia teacher education has kept up pace with the professional

demands. For more than 60 years now, Croatia has had access to qualified native

speakers who came to teach at Croatian universities through international organizations

and exchange of academic staff with English-speaking countries. In-service seminars for

teachers have been held on a regular basis involving teacher trainers from English-

speaking countries.

The quality of teaching is extremely difficult to tap into, but some empirical

studies provide insights into this area. In Hungary classroom observation projects

(Nikolov, 1999b, 2002b) and a nation-wide survey into the frequency of typical

classroom activities in primary and secondary schools revealed that teachers most often

apply techniques of the audio-lingual and grammar-translation method both in English

and German classes (Nikolov & Csapó, 2002; Nikolov, 2003a) and school achievements

tend to be generally low. Another important feature of foreign language education relates

to major differences between schools, classes and language learning groups within

schools (Csapó, 2001; Józsa, 2003; Nikolov, 2003b). Schools try to respond to parental

pressure and stream children from an early age in lower primary grades when foreign

language programs are launched. Learners with highest achievements in math, the mother

tongue, and other subjects are placed in groups starting earlier than mandatory, while the

rest of the students either start in grade 4 (age 10), the age prescribed in the National Core

Curriculum, or are offered German. Thus, no wonder Medgyes and Öveges (forthcoming)

are right stating that „much less has been achieved since the change of regime than has

been hoped for: the average Hungarian youth still does not speak foreign languages”.

These findings need to be interpreted in the light of data reflecting exceptionally

favorable attitudes towards and motivation to learn foreign languages, most importantly

English and German (e.g., Csizér, Dörnyei & Nyilasi, 1999; Nikolov, 2003a). Also,

Hungarian schools offer language instruction in groups: classes are divided into groups

for language lessons.

8

Students in Croatia benefit not only from better teaching, but they are also

exposed to a wide range of subtitled TV programs, while on Hungarian television all

intrinsically motivating programs are dubbed and few students benefit from international

channels (Nikolov & Csapó, 2002). Croats have been found to make good use of the

English language input they are surrounded with on a daily basis (Mihaljevic & Geld,

2003).

The Croatian national curriculum of modern languages has been stable over a

long period, whereas in Hungary the National Core Curriculum has gone through various

changes over the last 15 years reflecting functions of an interface between political and

professional preferences and exerting little influence whatsoever on classroom processes

(Medgyes & Nikolov, 2002). In Croatian schools instruction does not enhance individual

differences but helps students develop along more similar lines than in Hungary where

learners are streamed into groups according to their abilities from an early age (Andor,

2000; Csapó, 1998, 2001; Halász & Lannert, 2001, 2003).

Less variety in the TEFL curriculum in Croatia makes it easier for schools to offer

courses matching learners’ levels, while in Hungary a range of weekly hours, starting

year and other factors cause greater differences between groups and schools and make

transfer and continuity difficult (Vágó, 2000). Learners are often put in false beginner

groups and this practice may result in their loss of motivation (Nikolov, 2001).

Previous large-scale surveys of foreign language levels in Hungary

Hungarian students’ foreign language skills were assessed to monitor the levels

and efficiency of foreign language education in state schools in 2000 and 2002 (Csapó &

Nikolov, 2002; Nikolov & Csapó, 2002) and in 2003 (Józsa, 2003; Nikolov, 2003b)

similarly to major Hungarian educational research projects in other school subjects

(Csapó, 1998; 2002). Besides language skills tests, questionnaires were also administered

to the participants to collect data on their social and language learning background and

plans, attitudes, motivation, and classroom activities. To estimate their general cognitive

abilities, a standardized inductive reasoning skill test was used (Csapó, 1998).

Representative samples of Hungarian students participated in three projects

assessing their English and German language skills in 2000, 2002 and 2003. The units for

9

sampling and data collection involved school classes of grades 6, 8, 10 and 12 from

approximately 300 state schools, including close to 50,000 FL learners. Some findings of

these large-scale assessment projects are relevant for the present study: (1) students’

language performances showed a lot of variation in all cohorts: learners in some groups

achieved top scores, while other groups performed on extremely low levels; (2) learners

of English outperformed learners of German; (3) the best performances were found in

listening, partly, because the tasks were pitched at a lower level in light of classroom

observation data: listening comprehension is the least frequently developed skill in

Hungarian schools; (4) students’ achievements were the lowest in writing; (5) all

learners’ language learning attitudes, motivation, and long term plans with language

study were extremely positive, but slightly more favorable for English than for German;

(6) modest to strong correlations were found between the size of settlement, learners’

socio-economic status, cognitive skills, weekly language classes, and type of school for

secondary schools. These results indicate that in an educational context like Hungary,

where access to the examined target languages is mostly limited to the classroom,

students’ socioeconomic status and developmental level of cognitive skills strongly

influence their choice of and achievements in a foreign language. Similar findings

characterize achievements in other school subjects (Andor, 2000; Csapó, 1998; 2002).

Previous surveys of foreign language levels in Croatia

An attempt to look into the FL levels as well as into ways of making teaching and

learning more efficient was made by Filipović and associates (Filipović, 1971, 1974) as

part of Yugoslav Serbo-Croatian – English Contrastive Project (YSCECP). Results of the

project were used as the basis for designing new teaching materials (Vilke, 1975).

In the 1970s an early FL learning project was implemented in order to find out

what levels school children could achieve if they started in 2nd grade (age eight) (Vilke,

1976, 1976a, 1978). A more comprehensive project, initiated in the ‘90s, offered

research-based insights into the ways and conditions necessary to achieve useful levels of

FL competence by the end of primary education (Vilke & Vrhovac, 1993, 1995).

Although the project was sponsored by ministries, the results were ignored when

10

compulsory FL instruction was introduced in 1st grade (age six) on a national scale in

2003.

In 2003, as part of a national project aiming to find out about the levels Croatian

learners of English achieved by the end of primary and secondary education, more than

2,000 learners’ competence was tested using communicative tests developed and

validated in Hungary in 2002 and piloted in Croatia to check their validity in the new

context. A smaller sample of learners of German was administered the same type of

measuring instruments. These tests were used because they were considered valid for the

Croatian context and to allow comparisons between the two countries.

The findings of a comparison between Croatian and Hungarian 8th graders’ results

in 2002 indicated that while there were no significant differences in the listening

comprehension skills (because of a ceiling effect), Croatian learners were significantly

better than their Hungarian counterparts at reading comprehension and writing skills. A

comparison of the results obtained by learners of English and learners of German

revealed that English learners were better at reading comprehension and writing, while in

listening there were differences only in the secondary groups (learners of English being

better than learners of German) (Mihaljevic Djigunovic, in preparation).

The study

The present study is innovative in several ways: (1) Participants are learners of

neighboring countries in Central-Eastern Europe where language learning traditions and

the educational context share a lot of similarities, but there are also significant differences

which may impact on the outcomes. (2) The construct of language knowledge was the

same for first languages (Croatian and Hungarian) and English as a foreign language. (3)

The English proficiency tests were piloted and used with a nationally representative

sample in Hungary in 2002 and later with a smaller sample in Croatia. In the present

study they were used with new cohorts of Croatian and Hungarian learners. (4) A new

instrument was developed to tap into participants’ pragmatic competence both in

Croatian and Hungarian and EFL, and a new version of an EFL oral test was also

administered to a subsample in both countries. (5) The same questionnaire on background

11

was used in both contexts. (6) Schools were chosen from two neighboring regions and

cities in Croatia and Hungary: Slavonia including Osijek and Baranya county including

Pécs.

Research questions

The aim is to find answers to the following research questions:

1. How do Croatian and Hungarian 8th graders’ performances compare on EFL tests?

2. How do Croatian and Hungarian students’ performances compare on EFL tests by

groups and within groups?

3. How do performances compare according to length of study, number of weekly

classes, and size of group?

4. What is the relationship between Croatian and Hungarian students’ achievements

on the L1 and L2 tests? To what extent do L1 test scores predict L2 test scores?

PPaarrttiicciippaannttss

A total of 717 8th graders (distributed in 39 groups) participated in the study in Slavonia

and Osijek in Croatia and in Baranya and Pécs in Hungary (see Table 1).

----------------------------------------- Insert Table 1 about here

-----------------------------------------

The two samples (one in Hungary and one in Croatia) were drawn from year 8

(age 14) primary school EFL learners. This particular age group was chosen for two

reasons: (1) In both countries year 8 represents the end of primary school (and

compulsory education in Croatia) and it is the time when learners either end their formal

education or transfer to a secondary school (as the majority do). (2) By year 8 all learners

have had several years of EFL learning, regardless of when they started, and are able to

perform on proficiency measures. The participants were year 8 learners who represented

small village, small town and big town school EFL learners. These types of schools were

considered relevant from the point of view of the socio-educational context and, also,

quality of instruction.

12


-----------------------------------------

Instruments

EFL tests For English as a foreign language a battery of tests consisting of two test booklets

(one on reading comprehension and one on listening comprehension, writing and

pragmatics) and a speaking test were used. The tests in the two booklets had been

designed and validated first in Hungary in 2002 and later in Croatia. The speaking test

was adapted from a validated speaking test for Hungarian EFL learners for the purpose of

this project. All the tests were meaning focused and were based on the CEFR levels the

EFL learners were expected to reach at end of year 8 (A1 and lower band of A2 level).

The structure and content of the two test booklets are presented in Table 3.


-----------------------------------------

The topics as well as task types were familiar to the participants: they were not

built on or borrowed from any particular teaching material but they were highly similar to

the ones in the course materials they used in school. The vocabulary and structures of the

texts were expected to be on or a bit beyond the level of the target population. The

estimated level was hypothesized to cover a relatively wide range (A1 and lower band of

A2) as EFL learners in year 8, in both countries, include learners who started learning

English at the mandatory start in year 4 and those who had started earlier in one to three

classes per week. The length of the texts in the items ranged from a word, an expression,

or a sentence, to a short passage. All tasks focused on meaning, not form, and reflected

the achievement targets defined in the national curricula for these age groups in both

countries. The texts were authentic, except for the listening tasks, where scripted

materials were used. The reading booklet was produced in two versions: the sequence of

the tasks was different, but the actual tasks were identical. The rubrics were given in

English and all listening and reading tasks started with an example to ensure that even if

the instructions were slightly beyond test-takers or they skipped them, they could

understand what to do.

13

In Reading task 1 participants were supposed to match 10 basic vocabulary items

with ten definitions; for example, «hospital-a place where you visit someone sick or ill»;

«postman-a person who delivers letters». In Reading task 2, there were 10 public notices

to be matched with the clause/sentence that explains its meaning: for example, «use

before 01.02.2002-after this date you must not eat this food»; «return books here-this is

the library desk». In Reading task 3 participants had to read a short interview from a

British youth magazine and match nine questions and answers: for example, «What do

you do?-I go to Pinewood School.»; «Your favourite food?-Anything sweet, I love cakes

and ice cream.». Reading task 4 contained eight quiz questions that had to be matched

with appropriate answers: for example, «Why do some flowers smell sweet?-It's a way of

attracting bees, flies, butterflies and other insects»; «Which is the biggest bird of prey?-

The Andean condor measures up to 110 cm long and weighs up to 12 kg. Its wingspan is

over 3 metres.». In Reading task 5 participants were to match nine advertisements with

missing words.

Listening task 1 required participants to listen to a conversation between two

friends describing a picture and fill in the appropriate numbers in the table provided for

the persons in the picture next to the names. Although the listening text was scripted, the

dialogue and the situation sounded highly authentic. In Listening task 2 participants

listened to 10 short texts and worked on multiple choice items: with seven texts they were

supposed to decide where the dialogue took place, and the rest resembled standard

listening comprehension questions.

In the Writing task participants had two pictures that looked the same but

included ten small differences. They were told in the instructions which parts of the

picture the ten differences concerned, so they were not tested for their perception of the

differences.

The Pragmatics test contained twelve examples of everyday situations that were

described verbally and illustrated by a picture. Participants were to answer a question

about each situation by choosing the appropriate answer from the multiple choice items

offered. For example:

You’ve had dinner with an English family and you are now leaving because it’s

very late in the evening. What do you say to them?

14

A ”Thank you for the food. It was really good.”

B ”Thank you for inviting me. I really enjoyed the dinner.”

C ”It was a good dinner. Good evening.”

D ”Enjoy your meal.”

The Speaking test was administered to six, randomly chosen, students from each

school. The interlocutors were trained prior to going to the schools. The interviews were

carried out individually and all were audio taped. The test lasted for up to 15 minutes and

was strictly structured. It consisted of three tasks. Task 1 lasted 2-3 minutes and consisted

of the interlocutor asking nine questions: the first three were general questions (What's

your name? Could you spell your name, please? How old are you?), the other six could

be selected from the remaining nine. In Task 2 participants were first to choose one of six

pictures spread out on the table, describe it and explain the similarities and differences

between the scene in the picture (e.g. a busy street or messy kitchen) and the same place

in their own life. The task lasted 4-5 minutes. In Task 3, which also lasted 4-5 minutes,

participants were to choose two of six situations and act them out with the interlocutor.

For example:

Your friend is coming to visit you. Give him/her directions from the nearest

station or bus stop to your home.

You would like to cook something nice with your friend. Discuss what you like or

dislike and why.

In the first situation the interlocutor initiated conversation, while in the second

one the interviewee was invited to initiate it.

L1 tests

L1 tests booklets (both Hungarian and Croatian) included tasks in listening

comprehension, reading comprehension, writing, and pragmatics. These were new tests

designed and piloted on a similar population in both countries specifically for the purpose

of this project. The reading and listening comprehension tests followed the same format

15

in the two languages, but included different authentic input texts. The task types were

also the same and included either multiple choice items or short answers to be supplied

by participants.

In the Listening tests, both in Hungarian and Croatian, there were three tasks.

These were recordings of authentic national radio programs that 14-year-olds would be

expected to listen to and find intrinsically motivating. The task types used in the tests

tapped close listening as well as skimming.

The Reading tests included five tasks. In terms of content the tasks included

popular science texts, tables, and news from daily papers, texts from encyclopedias and

literary texts. Two tasks were the same: they were Hungarian tasks translated into

Croatian. The reading subskills tapped were skimming, scanning as well as intensive

reading.

The writing tests in both languages included the same two task types. Writing task

1 was the same in content too: it required participants to write about their opinion on

allowing learners to draw graffiti on their local school wall. In Writing task 2 participants

were asked to write a letter to a travel agent concerning a place where the participant and

three friends could stay during holidays. The content was culturally appropriate for the

two L1 writing tests as the names of the agents and places to stay were specific to the

Croatian and Hungarian contexts.

The pragmatics test included one task with six multiple choice items. The same

task was used in both countries, as from the socio-cultural point of view, they are highly

similar. The items concerned such examples as asking politely for a bill in a restaurant,

and reacting to friends being late for the cinema.

Questionnaire

The same questionnaire was administered in the participants’ mother tongue.

Besides detailed background data, participants were asked about the time they started

studying English and how many classes they had in a week.

Procedures

16

The paper and pencil tests were administered to students in classroom-size groups

in May, 2004. Participants had 45-minute class session for the reading booklets in each

language, 45-minute class sessions for listening and writing tasks in each language,

whereas a separate session was devoted to filling in the questionnaires. All tests and

questionnaires were coded and assessed centrally during the summer of 2004. Schools

were given their coded results in September. Oral tests, observations and interviews with

EFL teachers were conducted during the week or the week following test administration.

Six students were randomly chosen for the oral test; these tests were conducted with

external interlocutors.

As is evident from the descriptions of the instruments used to measure proficiency

in English and in L1, they included tasks that were simple to score and assess (e.g.

multiple choice items), whereas others required some negotiation (short answers) and the

construction of assessment tools and sophisticated training (writing and speaking tasks).

Assessment of short answers (in L1 reading and listening comprehension tasks)

was tackled in the following way: (1) test designers put down expected answers for each

item; (2) they read all papers and collected answers not matching their expectations; (3)

they consulted with colleagues to decide what to include in the final key.

For the writing tasks (one in English and two in the L1) three separate assessment

scales were constructed along four criteria: task achievement, vocabulary,

grammar/accuracy and text with scales of four bands each.

Assessment of speaking performance was done by means of a specially designed

assessment scale constructed along the following criteria: task achievement, vocabulary,

accuracy and fluency, pronunciation and intonation with a scale including five bands.

The assessors of both writing and speaking were trained. Since such training has

to focus on the actual tasks, four sets of training were conducted. Length of the training

depended on how much time the assessors needed to standardize their criteria (between

three to five hours). Standardization across the two countries was achieved in training

workshops with a trainer from Hungary. As for the L1 writings, one of the Croatian

assessors was a bilingual speaker of Hungarian. She had access to sample papers in

Hungarian with the actual scores given by two Hungarian assessors after negotiations

with one another and one member of the Croatian team.

17

All participating schools received coded statistical data from all schools and

detailed information on their students’ performances in November 2004.

RReessuullttss

Summary statistics for the participating students’ L2 performance on various

skills and for two total scores are included in Table 4. L2TOTAL includes data on

listening, reading and writing, whereas L2TOTALS includes results on the speaking test

in addition to the first three skills. As the data in Table 4 show, Croatian students

outperformed their Hungarian counterparts in all skill areas and, consequently, their total

scores are also higher. To test the hypothesis that these differences are statistically

significant, we applied independent samples t-tests to the data. The results of this analysis

are presented in Table 5. As can be seen, the difference between Croatian students’

English listening and reading scores turned out to be significantly higher than those of the

Hungarian students. The same was true for the two totals (L2TOTAL and L2TOTALS)

as well.

----------------------------------------- Insert Tables 4 and 5 about here

-----------------------------------------

To see how the performances of L2 groups in the two countries differed from one

another on the total English score variable, we first computed the group means for every

group in the two countries. Descriptive statistics for the obtained group means are

displayed in Table 6. The mean for the lowest scoring group was 30.92 points for

Hungary and 51.36 points for Croatia, while the mean for the highest scoring group was

84.33 points for Hungary and 83.82 points for Croatia. These results indicate that the

range of mean scores between the lowest and highest scoring groups was much higher for

Hungary (53.41 points) than Croatia (32.46 points). As a result, the variance of mean

group scores (Hungary: s2 = 423.29, Croatia: s2 = 67.81) was also significantly higher for

Hungary (F = 36.99, p < .001), showing that there is significantly more between-groups

variation in students’ performance in Hungary than in Croatia.


18

-----------------------------------------

In order to see whether within-group variation of students’ English performance

was also different for the two countries, we computed the within-group variances for each

group in both countries, which we followed up with an independent samples t-test testing

the hypothesis that the mean within-group variance was different for Hungary and

Croatia. The test indicated that no such difference existed, suggesting that students

performances within language groups were equally varied in the two countries.

We performed a multiple regression analysis to investigate the relationship

between L1 listening, L1 reading, and L1 writing measures as the predictor variables and

English total scores as the dependent variable. Two models were built, one for each

country. All the predictor variables were entered into the model simultaneously. The

standardized beta coefficients for the predictor variables in the resulting models along

with the corresponding t statistics are displayed in Table 7.


-----------------------------------------

For the Hungarian cohort, all the variables in the regression equation significantly

contributed to the prediction of students’ English test scores with L1 listening being the

best predictor (β = 0.408), followed by L1 writing (β = 0.284) and L1 reading (β =

0.194). In contrast, L1 reading (β = 0.625) turned out to be the only significant predictor

of L2 performance in the Croatian model. The total variance in L2 performance explained

by students’ L1 performance was consequently higher for Hungary (55%) than for

Croatia (32%).

We also collected data concerning the students’ age at which they started learning

English, the number of weekly classes, and the size of groups in which they study

English. To compare the relationship between starting age and L2 performance, we first

computed Pearson product-moment correlation coefficients between each L2 measure

and starting age separately for Hungarian and Croatian students. The resulting

correlations are displayed in Table 8.


19

-----------------------------------------

As data in Table 8 indicate, all coefficients turned out to be negative and significant

at the p < .01 level of significance, indicating that an earlier start might be beneficial in

both countries. Another general observation is that the absolute values of the coefficients

for the Hungarian subsample are all higher than the corresponding Croatian coefficients.

To examine if the differences between the coefficients are statistically significant, we

carried out a series of analyses of covariance (ANCOVAs). To do this, first we

standardized the variables that we correlated with each other. Using an ANCOVA, it is

possible to test the homogeneity of two regression slopes, which in the special case of

using standardized variables is equivalent to testing the equality of two correlations

between two variables for the two countries. The results of the ANCOVAs supported the

hypothesis that in Hungary there is a stronger negative relationship between learners’

starting age and all aspects of their L2 performance measured in the study when

compared to the situation in Croatia, since all results turned out to be significant at the p

< 0.01 probability level. That is, starting earlier seems to be more desirable in Hungary

than in Croatia.

As for the number of weekly classes, the average of weekly classes was 2.94 in

Croatia, and 3.52 in Hungary, whereas there were 22 and 14 students on average in

Croatian and Hungarian groups respectively.

Discussion

As the overall results indicate, Croatian students performed on a significantly

higher level in EFL than their Hungarian peers. This is a remarkable outcome of our

study in the light of three key factors. The first is the weekly number of classes indicating

frequency: although Hungarian learners get more English instruction and a strong

relationship was found within the Hungarian learners’ cohort between hours of

instruction per week and performance on the EFL tests, Croatian students achieved better

results in only three classes a week. The results of Hungarian students coincide with

findings in a nation-wide survey involving learners of English in years 6 and 10 where

higher weekly number of classes resulted in higher achievements (Józsa, 2003; Nikolov,

20

2003b, with one exception to this pattern: learners of German in the highest weekly

classes performed on a lower level than their peers in less intensive courses). Also,

Alderson (2000b) found that the relationship between weekly classes and outcomes was

more complex for a convenience sample of secondary-school EFL students.

The other indicator is the number of students in a group. Croatian learners study

English in whole classes, while Hungarian students learn in small groups; however, fewer

students in groups (an average of 14 versus 22) do not necessarily ensure better

achievements. Finally, an early start does not seem to guarantee higher achievements in

year 8, as the earlier starter Hungarian learners did not outperform their Croatian

counterparts, though learners with more years of instruction achieved better results

among the Hungarian participants. Therefore, the differences must be caused by the

quality of English language teaching and influences outside the classroom in the two

educational contexts.

An important finding may shed light on a different aspect of teaching and

learners’ development. As the results show, more variation was found between groups in

Hungary than in Croatia, but there was no significant difference within groups in either

context. This phenomenon is most probably a result of streaming in Hungarian schools

into ability groups (Andor, 2000; Csapó, 1998, 2002), while this practice is not typical in

Croatian schools. It must be borne in mind that most frequently schools skim classes in

the lower primary years and place more able young learners in the early groups, thus it is

impossible to tell whether it is the earlier start or learners’ cognitive abilities, or these two

in interaction, ensuring more favorable outcomes in year 8. How the time of start factor

influences outcomes is presently unclear, but it is certain that the assumption underlying

streaming - allow fast learners to develop at a higher pace and provide slow learners with

more support - does not seem to work for the latter group in Hungary. Less able learners

are often left behind. Therefore, the explanation of the strong relationship between

weekly classes and better achievements is that more able students are streamed into more

intensive groups, thus allowing them to develop even faster, while the less able are placed

in “normal” groups (Nikolov, 2004).

As in the present study learners’ aptitude was not measured, this assumption can

be supported by evidence from other studies. Csapó and Nikolov (2002) found that

21

inductive reasoning had an impact on FL achievement, similarly to Kiss and Nikolov’s

(2005) findings.

As to the extent to which Croatian and Hungarian learners’ achievements in L1

predict their achievements in English, in the case of Hungarian learners listening, reading

and writing test results in L1 significantly contribute to the prediction of EFL results,

while in the case of Croatian students only reading in L1 is a significant contributor.

Several explanations offer themselves: this finding may result from streaming Hungarian

learners into ability groups according to their achievements in L1, or the extensive use of

translation in EFL classes may contribute to a stronger interdependence between levels in

L1 and L2. As the findings of a nation-wide survey and classroom observations show

(Nikolov, 1999b; 2003b), translation into L1 is the most frequently applied classroom

technique of meaning making, practice, and assessment and it may influence outcomes.

As for the Croatian learners, it makes sense to explain the results by pointing out that

students with better L1 reading achievements can benefit more from exposure to L2

outside the classroom, most importantly from subtitled films freely available on

television.

Our study found evidence for the interdependence hypothesis, as a linear

relationship was found between L1 and L2 performances in both cohorts. Although

teachers and school officials often voice concerns reflecting a threshold hypothesis,

supposing that achievements in an L2 will be significantly lower below a certain level in

L1, our findings do not indicate such a threshold for these participants.

Conclusion and further research

Our study has indeed provided new insights into Croatian and Hungarian 8th

graders’ performances. Croatian students in year 8 performed on a higher level in EFL

than their Hungarian counterparts. Also, larger differences have been found between

groups’ performances in Hungary than in Croatia, whereas no significant differences

characterize within-group variations. As for the relationship between students’

performances in L1 and L2, the interdependence hypothesis has gained support in both

cohorts. The findings of our study suggest that factors usually assumed to be the keys to

success in FL learning, i.e. early start, more classes, and small groups, while extremely

22

important, do not guarantee higher achievements. The importance of these factors as well

as their interaction should be considered against a host of other variables, the most

important of which, in our opinion, are the quality of teaching and exposure to

comprehensible input in the target language. Exposure comprises both classroom

teaching and out of class contact with the FL. We believe that length, amount as well as

quality of exposure to the FL need to be taken into consideration. While there are

instruments for measuring the quality of teaching and the length and amount of FL

classroom teaching are easy to measure, the methodology for a systematic study of out of

class exposure is yet to be developed. In order to see the real impact of these variables

studies that would include them are warranted.

Further research is necessary to find out about the relationship between learners’

attitudes and motivation and achievements on the English tests; between their

performances on EFL proficiency tests and the observed classroom practices in the two

countries – not discussed in the present paper. It is particularly challenging to find out in

what ways classroom processes contribute to outcomes: how Hungarian and Croatian

teachers of English model the target language, what tasks they apply, how they scaffold

their learning, how much peer interaction takes place, and how they integrate learners’

extensive exposure to English outside the classroom. In other words, analyses of all data

collected in our larger project will hopefully provide insights into how the quality of

teaching may differ and, thus, offer learners in the two socio-educational contexts

different opportunities for language learning. Interview data with the teachers will

hopefully throw light on teachers’ professional beliefs about how learners acquire

English and benefit from classrooms tasks and outside class exposure, their motivation

and awareness of how their practices may relate to learning outcomes.

Further analyses are necessary to explore in what ways students’ speaking and

writing performances reflect developmental sequences in SLA studies; the relationship

between intended levels of difficulty of tasks in both L1 and L2 and students’

performances on them and how the CEFR (2001) ‘can do statements’ and the actual

spoken and written texts compare to one another. A different avenue of explorations into

the written and oral products would need to find answers to the question how linguistic

23

distance (Croatian – English; Hungarian – English), often hypothesized to influence

outcomes, interacts with other variables.

Finally, it is necessary to overview the implications of our comparative study for

teacher education, so that teachers may benefit from a better understanding of the

relationships between the wider social context and teachers’ and students’ classroom

behavior, teachers’ views expressed in the interviews, and students’ achievements on

tests.

24

References

Alderson, J. C. (1984). Reading in a foreign language: A reading problem or a language

problem (pp. 122-135). In J. C. Alderson & U. H. Urquhart (Eds.), Reading in a

foreign language. New York: Longman.

Alderson, J. C. (2000a). Assessing reading. Cambridge: Cambridge University Press.

Alderson, J. C. (2000b). Exploding myths: does the number of hours per week matter? In:

J. C. Alderson, E. Nagy & E. Öveges (Eds.), English language education in

Hungary. Part II: examining Hungarian learners' achievements in English.

Retrieved on 19 June 2005 from http://www.examsreform.hu/Pages/ELE2.html

Andor, M., (2000). A nyelvtudás szociális háttere. [Social background to language

knowledge] Educatio, 9, 4, 717-728.

Bachman, L. & Palmer, A. (1996). Language testing in practice. Oxford: Oxford

University Press.

Baker, C. (1996). Foundations of bilingual education and bilingualism. Clevedon, Avon:

Multilingual Matters.

Brantmeier, C. (2004). Statistical procedures for research on L2 reading comprehension:

an examination of ANOVA and Regression Models. Reading in a Foreign

Language, 16, 2, 51-69.

Cameron, L. (2000). Teaching languages to young learners. Cambridge: Cambridge

University Press.

Common European framework of reference for languages: learning, teaching,

assessment. (2001). Cambridge: Cambridge University Press.

Csapó, B. (1998). Az iskolai tudás. [School knowledge] Budapest: Osiris.

Csapó, B. (2001). A nyelvtanulást és a nyelvtudást befolyásoló tényezők. [Variables

influencing language learning and proficiency] Iskolakultúra, 1, 8, 25-35.

Csapó, B. (2002). Az iskolai műveltség. [School literacy] Budapest: Osiris.

Csapó, B. & Nikolov, M. (2002). The relationship between students’ foreign language

achievement and general thinking skills. American Educational Research

Association Annual Meeting, New Orleans, U. S. A. 1-5 April, 2002.

Csizér, K. Dörnyei, Z. & Nyilasi, E. (1999). Az általános iskolások nyelvtanulási

attitűdjei és nyelvválasztása Magyarországon. Az orosztól az angolig? [Primary

25

school students’ language learning attitudes and language choices: from Russian to

English?] Magyar Pedagógia, 99, 2, 193-204.

Cummins, J. (1991). Interdependence of first- and second-language proficiency in

bilingual children (pp. 70-89). In E. Bialystok (Ed.), Language processing in

bilingual children. Cambridge: Cambridge University Press.

Cummins, J. (2000). Language, power and pedagogy: bilingual children in the crossfire.

Clevedon, Avon: Multilingual Matters.

Curtain, H., & Pesola, A. (1994). Languages and children: Making the match. White

Plains, N.Y.: Longman.

Dörnyei, Z. (2001). Teaching and researching motivation. London: Longman.

Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University

Press.

Europeans and languages: A Eurobarometer special survey. (2001). European

Commission, Directorate-General for Education and Culture, Brussels. Retrieved on

20 May 2005 from http://europa.eu.int/comm/dgs/education_culture/index_en.htm

Filipović, R. (1971). YSCEP Pedagogical Materials 1. Zagreb: Institute of Linguistics,

University of Zagreb.

Filipović, R. (1974). YSCEP Pedagogical Materials 2. Zagreb: Institute of Linguistics,

University of Zagreb.

Genese, F. (1979). Acquisition of reading skills in immersion programs. Foreign

Language Annals, 12, 71-77.

Halász, G. & Lannert, J. (1998). Jelentés a magyar közoktatásról. [Report on Hungarian

education] Budapest: Országos Közoktatási Intézet, Budapest.





Johnstone, R. (2000). Context-sensitive assessment of modern languages in primary

(elementary) and early secondary education: Scotland and the European experience.

Language Testing, 17, 2, 123-143.

26

Józsa, K. (2003). Az idegen nyelvi készségek fejlettsége angol és német nyelvből a 6. és

10. évfolyamon a 2002/2003-as tanévben. Függelék. [Levels of performances in

English and German in year 6 and 10. Appendixes] Budapest: OKÉV.

Lee, J., & Schallert, D. L. (1997). The relative contribution of L2 language proficiency

and L1 reading ability to L2 reading performance. TESOL Quarterly, 31, 4, 713-

739.

Kiss, Cs., & Nikolov, M. (2005). Developing, piloting, and validating an instrument to

measure young learners’ aptitude. Language Learning, 55, 99-150.

Medgyes, P., & Nikolov, M. (2002). Curriculum development: The interface between political

and professional decisions (pp. 195-206). In R. B. Kaplan (Ed.), The Oxford handbook of

applied linguistics. Oxford: Oxford University Press.

Medgyes, P., & Öveges, E. (forthcoming). Paved with good intentions: Foreign language policy

in Hungary. TESOL Matters.

Mihaljevic Djigunovic, J. (in preparation). EFL or ELF? Focus on English in Croatia.

Mihaljevic, J., & Geld, R. (2003). English in Croatia today: Opportunities for incidental

vocabulary acquisition. Studia Romanica et Anglica Zagrabiensia, XLVII-XLVIII, 335-352.

Moon, J. (2000). Children learning English. Oxford: Macmillan.

National Census 2001, Summary Data, Volume 1. (2002). Budapest: Central Statistical

Office.

Nikolov, M. (1999a). The socio-educational and sociolinguistic context of the

examination reform (pp. 7-20). In H. Fekete, É. Major, & M. Nikolov (Eds.),

English language education in Hungary: A baseline study. Budapest: British

Council.

Nikolov, M. (1999b). Classroom observation project (pp. 221-246). In H. Fekete, É.

Major, & M. Nikolov (Eds.), English language education in Hungary: A baseline

study. Budapest: British Council.

Nikolov, M., (2000). Teaching foreign languages to young learners in Hungary (pp. 29-

40). In M. Nikolov & H. Curtain (Eds.), An early start: Young learners and modern

languages in Europe and beyond. Strasbourg: Council of Europe.

27

Nikolov, M. (2001). A study of unsuccessful language learners (pp. 149-170). In Z.

Dörnyei & R. Schmidt (Eds.), Motivation and second language acquisition.

Honolulu: The University of Hawaii.

Nikolov, M. (2002a). Issues in English language education. Bern: Peter Lang AG.

Nikolov, M. (2002b). What do teachers of young learners claim and do? An empirical

study of their claims and practices. Dijete i Jezik Danas/Child and Language

Today. 2nd International Conference, Osijek, Croatia, 21 November, 2002.

Nikolov, M. (2003a). Angolul és németül tanuló diákok nyelvtanulási attitűdje és

motivációja. [Attitudes and motivation of learners of English and German]

Iskolakultúra, 13, 8, 61-73.

Nikolov, M. (2003b). Az idegen nyelvi készségek fejlettsége angol és német nyelvből a

6. és 10. évfolyamon a 2002/2003-as tanévben. [Levels of performances in English

and German in year 6 and 10] Budapest: OKÉV.

Nikolov, M. (2003c). Az idegennyelv-tanítás megújulásának hatásai. [The effect of

innovation in modern language teaching] Új Pedagógiai Szemle, LII, 3, 46-57.

Nikolov, M. (2004). Social and classroom dimensions of young learners’ foreign

language learning. Issues in teaching, learning and assessment of L2 young

learners. University of Cambridge, Cambridge, U.K. 28-29 May 2004.

Nikolov, M., & Csapó, B. (2002). Twelve-year-olds’ attitudes towards classroom

activities and their performances on tests of English and German as a foreign

language. American Association of Applied Linguists Annual Conference, Salt Lake

City, U. S. A. 6-9 April, 2002.

Skehan, P. (1998). A cognitive approach to language learning. Oxford: Oxford University

Press.

Sparks, R., & Ganschow, L. (2001). Aptitude for learning a foreign language. Annual

Review of Applied Linguistics, 21, 90-111.

Vágó, I. (2000). Az oktatás tartalma. [Content of teaching] (pp. 169-238) In G. Halász &

J. Lannert (Eds.), Jelentés a magyar közoktatásról 2000. [Report on Hungarian

education] Budapest: Országos Közoktatási Intézet.

Verhoeven, L. (1991). Acquisition of biliteracy (pp.61-74). In H. N. Hulsijn & J. F.

Matter (Eds.), Reading in two languages. Amsterdam: AILA.

28

Vilke, M. (1975). On compiling pedagogical materials (pp. 63-77). In R. Filipovic (Ed.),

YSCEP, Studies 6. Zagreb: Institute of Linguistics, University of Zagreb.

Vilke, M. (1976). Implications of the age factor on the process of acquisition of an L2

(87-104). SRAZ XXI, Zagreb: University of Zagreb.

Vilke, M. (1978). A further inquiry into the process of acquisition of a second language at

an early age. AILA, Fifth International Congress of Applied Linguistics,

(Association Paper), USA.

Vilke, M., & Vrhovac, Y. (Eds.) (1993). Children and foreign languages I. Zagreb:

Faculty of Philosophy, University of Zagreb.

Vilke, M., & Vrhovac, Y. (Eds.) (1995). Children and foreign languages II. Zagreb:

Faculty of Philosophy, University of Zagreb.

29

Table 1: Participants in Croatia and Hungary

Croatia Hungary Total

Schools 10 10 20 Groups 21 18 39 Students 470 247 717 Teachers 12 13 25

30

Table 2: Location of participating groups/classes in Croatia and Hungary

Croatia Hungary Total

City 8 15 23 Town 8 1 9 Village 5 2 7 Total 21 18 39

31

Table 3:

Structure and content of EFL test booklets

Skill Task Input text No. of items

Reading 1 Match word with appropriate sentence

Definitions of words 10

Reading 2 Match notice with meaning

Public notices and their meanings

10

Reading 3 Match question with answer

Interview from youth magazine 9

Reading 4 Match questions with answers

Quiz texts 8

Reading 5 Match advert with missing word

Advertisements 9

Listening 1 Multiple choice on pictures

Conversation (scripted) 10

Listening 2 Multiple choice on texts

Conversations (scripted) 10

Writing Describing differences

Describing similarities and differences in two pictures

Marking scheme

Pragmatics Multiple choice Dialogues in various language use situations

12

32

Table 4:

Summary statistics for students’ L2 performance in the two countries

COUNTRY N MEAN SD

L2LISTENING Hungary 216 17.62 3.43 Croatia 416 18.38 2.86

L2READING Hungary 231 26.94 11.67 Croatia 353 32.34 9.84

L2WRITING Hungary 216 17.15 11.05 Croatia 320 17.62 9.08

L2SPEAKING Hungary 35 29.09 14.21 Croatia 95 29.69 12.12

L2PRAGMATICS Hungary 230 7.18 2.68 Croatia 379 7.40 2.28

L2TOTAL Hungary 202 62.22 24.62 Croatia 247 71.49 16.26

L2TOTALS Hungary 25 89.00 39.01 Croatia 67 109.31 20.31 Notes: L2TOTAL = L2LISTENING + L2READING + L2WRITING; L2TOTALS = L2TOTAL + L2SPEAKING

33

Table 5:

Comparison of students’ L2 performance in the two countries (independent samples t-test)

t df P

L2LISTENING -2.809 373.806 .005* L2READING -5.815 432.793 .000* L2WRITING -0.517 399.634 .605 L2SPEAKING -0.225 53.282 .823 L2PRAGMATICS -1.025 425.066 .306 L2TOTAL -4.593 335.051 .000* L2TOTALS -2.481 28.992 .019* Notes: Equal variances not assumed; * = significant at the p < .05 level.

34

Table 6:

Descriptive statistics for group means of L2TOTAL scores in the two countries

HUNGARY (N=16) a

CROATIA (N=21)

Minimum 30.92 51.36 Maximum 84.33 83.82 Range 53.41 32.46 Mean 58.88 70.71 Variance 423.29 67.81 Notes: a Two groups in the Hungarian cohort did not take the English writing test, therefore, the total scores were only available for 16 Hungarian groups.

35

Table 7:

Multiple regression analysis of students' L1 and L2 performance

Multiple R Adjusted R2 β t p

HUNGARY 0.744 0.546 L1LISTENING 0.408 6.276 .000* (N=183) L1READING 0.194 3.012 .003* L1WRITING 0.284 4.627 .000* CROATIA 0.590 0.315 L1LISTENING -0.103 -0.950 .346 (N=64) L1READING 0.625 5.054 .000* L1WRITING -0.033 -0.275 .784 Notes: * = Significant at the p < .01 level.

36

Table 8:

Correlations between L2 performance and starting age

HUNGARY CROATIA

L2LISTENING -0.504 (200) -0.169 (412)

L2READING -0.592 (216) -0.259 (350)

L2WRITING -0.675 (199) -0.218 (317)

L2SPEAKING -0.810 (32) -0.285 (94)

L2PRAGMATICS -0.421 (216) -0.192 (376)

L2TOTAL -0.656 (187) -0.254 (246)

L2TOTALS -0.827 (23) -0.326 (67) Notes: All coefficients are significant at the p < .01 level. N sizes are given in parenthesis.

a comparative study of croatian and hungarian 8th graders ... · croatian is a slavic language,...

Documents